manual position committed

[keith preston]

>
>
> There's a "valid_until" gconf-key that needs to be set to a value higher
> than current time, or the backend won't emit the signals. I'm still open
> to other, more generic solutions.


I think the valid until, is an interesting idea.   I don't think it should
stop emitting the position after the valid time, but rather have an
significant increase in error.


Todo list:
> * Implement service_changed signal emission
> * Add gconf schemas
> * refactor (it's fairly rough right now, but I wanted it in so
>    Keith won't break my tree with more API changes :)
>
> Some observations about API changes:
> * civic location method/signal should have timestamp argument too?
>
* I assume timestamp is time of original data retrieval,
>    in case of manual backend that would be when user inputs the data?
> * Maybe current_position should actually work like "last_position":
>    always return the last valid position and let the client check if
>    timestamp is acceptable? If clients are interested in current data,
>    they should subscribe to signals anyway... The function should be
>    renamed though: "current" just doesn't describe this behaviour.


This is an interesting idea.   I believe this is what I intended with the
timestamps.   I believe the more we work with this, the more we find that
multiple backends will be needed to find a good location.   I believe
timestamps will help us make the 'geoclue-master' arbitrator much easier.

Keith Preston


_______________________________________________
> GeoClue mailing list
> GeoClue at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/geoclue
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freedesktop.org/archives/geoclue/attachments/20070815/c9113709/attachment.htm